GOALI: Quality Mining - A Novel Framework for Quality Monitoring and Control for Data-rich Manufacturing Systems

该补助金提供资金，通过结合计算机和信息科学原理，例如用于网络挖掘的原理，为制造系统开发新的质量工具。这些新设想的质量方法将利用数据挖掘技术，如关联，聚类和重要性，作为处理大量的过程和产品数据收集在当前复杂的制造系统，以便将数据转化为可用的知识的手段。这些新的质量工具将被开发用于存储、组织、分析、建模和可视化与制造系统相关的大型异构数据集，以实现持续的质量和可靠性量化和改进。这些技术将把当前的数据推送质量方法（通常在没有明确目的的情况下生成报告）转变为数据拉取系统，在该系统中，质量系统将反映对任务环境的认识，同时对故障（警报条件）或用户查询做出反应。这项工作包括与QMC LLC（一家数据管理公司，将实施开发的工具）、FARO（空间数据采集系统提供商）和福特汽车公司（将提供测试台，在工业环境中验证拟议的工具）的密切合作。 这项研究的结果将重新定义当前的质量控制方法，通过将低层次的数据转化为高内容的信息，从而更深入地了解制造和服务系统。这将允许更快、更准确的故障检测，从而显著提高质量。此外，有效处理数据丰富的制造系统的能力将刺激先进传感器和数据收集技术的开发和整合。这将产生基于异构数据管理的持续质量改进态度。

This grant provides funding to develop new quality tools for manufacturing systems through incorporating computer and information science principles, such as those used in web mining. These newly envisioned quality methods will make use of data mining techniques such as association, clustering, and significance as a means of handling large amounts of process and product data being collected in current complex manufacturing systems in order to transform the data into usable knowledge. These new quality tools will be developed to store, organize, analyze, model, and visualize large, heterogeneous data sets associated with manufacturing systems for continuous quality and reliability quantification and improvement. These techniques will transform the current data-push quality approach, where reports are often generated without a clear purpose, into a data pull system where the quality system will reflect an awareness of the task environment while reacting to a fault (alarm condition) or a user query. This work includes a strong collaboration with QMC LLC, a data management company who will implement the developed tools, FARO, provider of dimensional data acquisition systems, and Ford Motor Company, who will provide a test bed to validate the proposed tools in an industrial environment. The results of this research will redefine current quality control methods through transforming low-level data into high-content information leading to a deeper understanding of manufacturing and service systems. This will allow for quicker and more accurate failure detections, leading to a significant increase in quality. In addition, the ability to effectively handle data-rich manufacturing systems will stimulate the development and incorporation of advanced sensor and data collection techniques. This will produce a continuous quality improvement attitude based upon heterogeneous data management.